Since earley parcer finds all possible application variants for a token, can it parce text in parallel, unlike the usual parcer like stack-based, etc. You just need to modify the start of each parallel chunk of tokens, then while going backwards while constructing a table you combine and validat...